Error and sensitivity analysis scheme of a new data compression technique in estimation

Abstract

A new data compression technique of a new class was recently introduced. Although the rationale and the equations involved in the application of the technique were fully presented, an error analysis of the actual (true) estimation error was not provided. Due to the unconventional operations involved in the application of the technique, the standard error and sensitivity analysis routines could not be utilized. This paper presents a genuine error and sensitivity analysis scheme which handles the error analysis problem of the new data compression technique. The scheme is based on a reformulation of the data compression equations in a compact form. This, together with the proper sequencing of the operations involved in the execution of the data compression technique, makes the technique compatible with the standard practice of error and sensitivity analysis. Particular care is given to the analysis of biased estimates. An example involving data compression in an inertia! navigation system is introduced, and computer results are presented and discussed.